Scholarlink

<p>The Background Error Covariance Matrix (B) plays a crucial role in the Data Assimilation (DA) system, influencing the accuracy of numerical weather predictions. In this study, we investigate the impact of different B formulations on the sensitivity of model Analysis (ANA) and the forecast quality of intense precipitation events over Indian region during South West Monsoon season. Numerical DA experiments are conducted with 3DVar, 3DEnVar, and 4DEnVar techniques at a 30-km horizontal grid resolution. Four distinct B formulations are explored: (i) B-CLIM using the National Meteorological Center (NMC) method; (ii) B-MPCU by incorporating 45-member ensembles by varying microphysics (mp) and cumulus (cu) schemes; (iii) B-RAD by altering mp, cu, short and long-wave radiation schemes; and (iv) B-PERT by perturbing initial conditions while formulating 45-member ensembles. Initial Conditions (IC) and Boundary Conditions (BC) are derived from NCEP GFS at 6-hourly intervals and 0.25° <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\:\times\:\)</EquationSource> </InlineEquation> 0.25° spatial resolution. Surface, upper-air, and GPS radio occultation data are utilized in DA. The focus is on ten extreme rainfall events during the South West (SW) monsoon season from 2022 to 2018, associated with severe weather systems (Depression or more severe) formed over the Bay of Bengal (BOB) region. Comparative analyses, including model ANA against radio-sonde, METAR observations, and ERA-5 data and model forecast against the GPM 0.1°<InlineEquation ID="IEq2"> <EquationSource Format="TEX">\(\:\:\times\:\:\)</EquationSource> </InlineEquation>0.1° rainfall estimate, reveal that the Hybrid-3DEnVar technique with B-MPCU outperforms other approaches in all the ten cases. Our experiment results indicate that the incorporation of flow-dependent B, derived by varying mp and cu in the assimilation technique, enhances the forecast model’s performance for severe weather systems by improving ANA and predictions over Indian region during SW Monsoon season.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Formulation of background error covariance matrix in data assimilation and its impact on prediction of very heavy rainfall over Indian region during South West monsoon season

  • 期刊论文

  • 发表日期：

    2026年3月5日

查看全文
  • M. K. Chandragiri,
  • S. Dubey,
  • S. Baidya Roy,
  • J. P. George

摘要

The Background Error Covariance Matrix (B) plays a crucial role in the Data Assimilation (DA) system, influencing the accuracy of numerical weather predictions. In this study, we investigate the impact of different B formulations on the sensitivity of model Analysis (ANA) and the forecast quality of intense precipitation events over Indian region during South West Monsoon season. Numerical DA experiments are conducted with 3DVar, 3DEnVar, and 4DEnVar techniques at a 30-km horizontal grid resolution. Four distinct B formulations are explored: (i) B-CLIM using the National Meteorological Center (NMC) method; (ii) B-MPCU by incorporating 45-member ensembles by varying microphysics (mp) and cumulus (cu) schemes; (iii) B-RAD by altering mp, cu, short and long-wave radiation schemes; and (iv) B-PERT by perturbing initial conditions while formulating 45-member ensembles. Initial Conditions (IC) and Boundary Conditions (BC) are derived from NCEP GFS at 6-hourly intervals and 0.25° \(\:\times\:\) 0.25° spatial resolution. Surface, upper-air, and GPS radio occultation data are utilized in DA. The focus is on ten extreme rainfall events during the South West (SW) monsoon season from 2022 to 2018, associated with severe weather systems (Depression or more severe) formed over the Bay of Bengal (BOB) region. Comparative analyses, including model ANA against radio-sonde, METAR observations, and ERA-5 data and model forecast against the GPM 0.1° \(\:\:\times\:\:\) 0.1° rainfall estimate, reveal that the Hybrid-3DEnVar technique with B-MPCU outperforms other approaches in all the ten cases. Our experiment results indicate that the incorporation of flow-dependent B, derived by varying mp and cu in the assimilation technique, enhances the forecast model’s performance for severe weather systems by improving ANA and predictions over Indian region during SW Monsoon season.