Verifications of CPS for one-month prediction using its Hindcast(JMA/MRI-CPS4)

Hindcast Verification

• Bias map (Mean error map)
• Northern hemisphere map
• Global map
• Zonal mean map

• Hindcast maps for every initial date
• Northern hemisphere map
• Global map

• Verification score
• Time-series Circulation Index
• Verification Score Maps
• Variables to be Assessed: RAIN, Z500, T850, SLP, CHI200, PSI200, PSI850
• Diagonostic Measures:
• Anomaly Correlation(ACOR)
• Root Mean Squared Error(RMSE)

• Deterministic score Maps
• Variables to be Assessed: T2m, SST, RAIN
• Diagonostic Measures:
• Mean Square Skill Score(MSSS)

• Probabilistic score Diagrams
• Variables to be Assessed: T2m, SST, RAIN
• Diagonostic Measures:
• Reliability diagrams (Aggregated verification)
• Relative Operating Characteristics(ROC) curve (Aggregated verification)
• Event: Anomaly > 0, Below Normal, Near Normal, Above Normal, All
• The 3 categories(above-normal, near-normal, below-normal) are defined by standard deviations of the analyzed multi-years time series at each grid point.

• Probabilistic score Maps
• Variables to be Assessed: T2m, SST, RAIN
• Diagonostic Measures:
• Relative Operating Characteristics(ROC) areas (Grid point verification)
• Event: Anomaly > 0, Below Normal, Near Normal, Above Normal, All
• The 3 categories(above-normal, near-normal, below-normal) are defined by standard deviations of the analyzed multi-years time series at each grid point.

Specifications of the hindcast

• Model: JMA/MRI-CPS4
• Ensemble size: 13
• Target period for verification: 1998-2020 (RAIN), 1991-2020 (the others)

Verification data

• Precipitation (RAIN) : GPCP (Global Precipitation Climatology Project) precipitation dataset (version 1.3; Adler et al. 2017)
• Atmospheric Analysis : JRA-3Q
• Sea surface temperature : MGDSST